Supporting Information Appendix Supplemental Literature Alcohol intolerance: Biological causes and consequences When alcohol is ingested into the human body, ethanol is metabolized producing acetaldehyde. This task is performed by alcohol dehydrogenase (ADH) in the liver. Subsequently, acetaldehyde is metabolized to acetate. This process is largely carried out by the enzyme aldehyde dehydrogenase 2 (ALDH2) (1). While acetate is harmless, acetaldehyde is toxic and causes facial flushing, nausea, and tachycardia, which are the main symptoms of the alcohol intolerance. The liver enzymes ADH and ALDH are polymorphic in humans (2). ADH2 and ALDH2 are common in East Asians, and the homozygote ADH21 is with high frequency mutated into the super active heterozygote ADH22, while the homozygote ALDH21 is mutated into the inactive heterozygote ALDH22, caused by the replacement of Glutamate (Glu) at position 487 with lysine (Lys) (3, 4). The high frequency of ADH22 in Asians causes their bodies to more quickly convert ethanol into acetaldehyde, while the ALDH22 dramatically decreases the speed at which acetaldehyde metabolizes into the harmless acetate. This is the reason for the high frequency of Asian flushing. Overall, about 36% of East Asians are intolerant to alcohol (5). It has been long known that alcohol intolerance influences drinking behavior and thus, for example, impacts the risk of developing alcoholism. Alcoholics exhibit significantly lower frequencies of the ADH22 and ALDH22 alleles than non-alcoholics (2). The genetic variation in both ADH and ALDH that causes the intolerance, as well as the resulting flushing reaction, accelerated heart rate, and nausea, act to discourage individuals from drinking. In turn, this reduces the risk of alcohol-related problems (6). Evidence of negative relation between alcohol intolerance and alcohol dependence has been found in mainland Chinese (7), Taiwan Chinese (2, 8), Japanese (9, 10) and Korean (11) populations. Alcohol intolerance increases the risk of several kinds of cancer in East Asians (12). Liver and esophageal cancers were among the first connected to the mutated S1 genotype of ADH2 and ALDH2. In particular, studies have shown that drinkers with the alcohol intolerance face higher risks for these two types of cancers (13, 14). Recently, gastric cancer has also been tied to the alcohol intolerance (15), as has colorectal (16) and pancreatic cancer (17).* Explanation of Hypotheses Hypothesis 1: Males with alcohol intolerance practice self-control in drinking environments more than tolerant males. Female self-control over drinking alcohol is not modulated by alcohol intolerance. Hypothesis 1 is supported by Hendershot et al. (23) which reports that males with alcohol intolerance develop drinking refusal self-efficacy, and suggested this is because they acquire more experience with refusing opportunities to drink. Because Chinese women are not typically expected to consume alcohol, and thus do not often need to refuse to drink, it is plausible that this finding does not extend to Chinese females. For example, one study found that females' average monthly alcohol consumption is significantly lower than that of males (6.6 ml vs 188.6 ml) and that, unlike males, there are no significant differences in drinking patterns (frequency, amounts consumed) between Chinese women who are and are not alcohol intolerant (7). Other studies report that 3% of Chinese men are alcohol dependent, but only 0.1% of Chinese women (24, 25). Hypothesis 2: Males intolerant to alcohol and regularly in drinking environments will develop generalized improvement in self-control in relation to otherwise identical tolerant males and intolerant males not regularly exposed to drinking environments. In particular, in relation to alcohol tolerant males and intolerant males not regularly exposed to drinking environments, intolerant males * It’s worth noting that humans’ social behavior, and in particular, risky behavior after drinking alcohol, has been studied extensively. Examples include impact on labor productivity (18), relation to risky sexual behavior (19, 20), and propensity to commit violent crimes (21). Recently, Corazzini et al.(22) studied economic behavior under the influence of alcohol. They focused on risk preferences, time perception and altruism after alcohol consumption. They found that alcohol intoxication did not affect risk attitudes, while it did increase impatience and leave people less generous. Our study differs from these in that we focus on whether people with the alcohol intolerance may have developed generalizable self-control strategies rather than investigating immediate effects of alcohol consumption on preferences or decisions. S2 regularly in drinking environments will demonstrate an improved ability to use willpower to resist selfish temptations. Hypothesis 2 requires that some males encounter, reasonably regularly, drinking environments. There is substantial evidence this is the case. In China, alcohol is commonly used by males as a part of business meetings, in order to maintain good relations both among and between supervisors and employees, and to promote camaraderie among colleagues (26). Hence, males with alcohol intolerance have an incentive to drink, but prefer to drink only a small amount. This substantial social-pressure requires them to develop skills to resist the impulse to drink, and thus alcohol-intolerant males are more likely to need to practice self-control systematically more than alcohol-tolerant males. Consequently, the strength model of self-control suggests alcohol-intolerant males should be better able to exercise willpower in order to avoid succumbing to selfish temptations. Hypothesis 3: Female self-control will be modulated neither by alcohol intolerance nor exposure to drinking environments. Hypothesis 3 again relies on significant gender differences in Chinese drinking culture. As noted gender differences in social pressure to drink are pronounced in China (24, 25, 27). While men may be encouraged to drink, women typically are not and indeed are often strongly discouraged from drinking (24, 25, 10). It follows that females generally need not exercise self-control in drinking environments, and thus their practice of self-control will not systematically differ between those who are and are not tolerant to alcohol, regardless of the frequency of their exposure to such environments. Consequently, the strength model of self-control does not predict differences between tolerant and intolerant females’ abilities to exercise willpower. S3 Supplemental Figures 1.0 Tolerant 0.9 Intolerant 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Proportion of Self-control Drinker 0.0 N=104 N=74 N=112 N=38 Male Female Fig. S1. Proportion of respondents indicating they use self-control in drinking environments by gender and alcohol intolerance. Those indicating no drinking experience are excluded. Differences are marginally statistically significant between tolerant and intolerant males (t=1.920, p=0.056, two-sided t-test; z=1.906, p=0.057, two-sided Mann-Whitney test). The difference between tolerant and intolerant females is not significant. Error bars represent 95% Confidence Interval. 1.0 Tolerant 0.9 Intolerant 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Frequency of High Payoff Reports 0.0 N=152 N=93 N=173 N=59 Male Female Fig. S2. Frequency of High Payoff Reported by Gender and Presence of Alcohol Intolerance. Frequency of high payoff numbers 3, 4 and 5 reported are statistically significantly different between tolerant and intolerant males (t=2.275, p=0.023, two sided t-test, same as below) but do not differ between tolerant and intolerant females (t=0.504, p=0.614). Frequency of high payoff numbers reported are also significantly different between tolerant females and tolerant males (t=3.764, p<0.001) but do not differ between intolerant females and intolerant males (t=0.284, p=0.776). Error bars represent 95% confidence intervals. S4 1.0 1.0 Tolerant Tolerant 0.9 Intolerant 0.9 Intolerant 0.8 0.8 0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 equency of High Payoff Reports Payoff of High equency Frequency of High Payoff Reports Payoff of High Frequency Fr 0.0 0.0 N=90 N=55 N=49 N=16 N=45 N=33 N=117 N=36 Male Female Male Female Regular Exposure Rare Exposure Fig. S3. Frequency of High Payoff Reported by Gender, Alcohol Intolerance and Exposure to Drinking Environments. Regular exposure means participating in a drinking environment one or more times per month; rare exposure is less than this. Frequency of high payoff numbers reported marginally significantly differ between tolerant and intolerant males with regular exposure (t=1.764, p=0.079), but not between tolerant and intolerant females with regular exposure (t=0.643, p=0.522). Among those with rare exposure, the frequency of high payoff numbers reported do not significantly differ between tolerant and intolerant males (t=1.236, p=0.220) nor between tolerant and intolerant females (t=0.553, p=0.580). Error bars represent 95% confidence intervals. S5 Supplemental Tables Table S1. A. Summary Statistics by Alcohol Intolerance and Gender Male Female Intolerant Tolerant P value Intolerant Tolerant P value Age 20.07 19.86 0.261 20.30 20.16 0.499 Demographics Grade 2.15 2.07 0.550 2.54 2.33 0.186 Altruism 4.88 5.75 0.096 6.03 6.85 0.131 Risk preference 12.11 12.55 0.493 11.28 10.90 0.556 Preferences Time preference1 27.33 27.84 0.774 30 26.16 0.223 Time preference2 60.59 61.48 0.309 59.5 60.24 0.409 Ambiguity aversion 0.65 0.61 0.755 0.72 0.57 0.251 Extraversion 24.90 25.93 0.163 23.98 25.59 0.053 Agreeableness 34.40 35.15 0.205 34.62 35.08 0.506 Big Five Conscientiousness 29.36 29.56 0.775 28.30 28.73 0.595 Neuroticism 23.97 23.20 0.277 24.28 24.20 0.928 Openness 34.17 33.86 0.716 33.86 34.38 0.598 Machiavelli 76.61 76.37 0.880 74.08 72.20 0.300 Self-Monitoring 9.37 9.46 0.843 9.32 9.05 0.659 Other Personalities Self-Esteem 24.23 23.59 0.457 25.32 24.17 0.226 CFC 65.15 66.42 0.433 64.61 64.50 0.950 Table S1.